On-line Service

We present a user-friendly web server, called SVMSignal, for signal peptides prediction of whole proteomes. The web service accepts in a query at most 10,000 sequences in FASTA format. In addition to prediction results, SVMSignal also shows calculated biochemical properties, including free energy, polarity, average volume and charge index of each residue, in a compact profile graph for visualization to allow users to conveniently inspect potential signal peptide structure such as n-region, h-region, and c-region. Users can download all information within the first 100 residues, including predictions in FASTA format and all of the profile values in tab-delimited text format and the graph of profiles.
The method incorporates contact propensities, various sequence, physico-chemical, and structural information in a two-level architecture using support vector machines (SVMs). In the first level, contact residues are predicted and their pairing relationship or connectivity is further predicted in the second level. The two-level method for contact prediction significantly reduces the computational cost and false positives compared to directly predicting the contact map based on our benchmarks. Helix-helix interactions can be inferred based on the residue contacts predicted by TMhit. Together with the contact propensities, this method can be used to gain insights into helix-packing in membrane proteins.
In this research, we present TMexpo, a method to predict rotational preferences of transmembrane helices to facilitate structural modeling. TMexpo first predicts lipid accessibility (the relative accessible surface area in lipid) by Support Vector Regression and predicts the classification of burial and exposed status of transmembrane helices (TMHs) by Support Vector Machine; and both models use evolutionary profiles, sequence conservation, helix insertion energy and biochemical properties as features. Then TMexpo calculates rotational angles of TMHs based on the predicted relative accessible surface area.
The Transmembrane Protein Helix-Packing Database (TMPad) is an integrated repository of experimentally determined structural folds derived from helix-helix interactions in alpha-helical membrane proteins. TMPad includes geometric descriptors of helix-helix interactions, topology, lipid accessibility, ligand and binding sites information. In addition, TMPad provides structural classification and visualization of the above structural features of TM helix-packing.